There are at least two choises to get OCR working in Obsidian for now:
I haven’t tried the Tesseract installation yet, but I’ve used the Taskbone service. At least it’s easy to setup and it works well, but I understand that in some cases it’s not desirable to send private data to a third party to process.
A core plugin would be welcome, I think. I took a quick glance on the Tesseract installation instructions, and it didn’t look too difficult, but I wish a plugin could download the Tesseract binary automatically. Now you need to download it manually on Windows & Linux. On Mac you can use a brew command.